|
|
Overview
Another research direction of mine is in integration/interoperability systems needed to hook up disparate systems into seamless environments. Systems on which I have worked include a Dynamic Backplane Architecture for large, integrated network modeling (e.g., combining wired and wireless models), and Runtime Infrastructure for the US Department of Defense’s High Level Architecture (HLA).
Network Modeling
An important issue in network simulation is in constructing rich scenarios with a variety of models. Traditionally, each network modeling package included its own set of validated models (e.g., of TCP & IEEE 802.11), but it was infeasible to combine models across different simulators for experimenting with complex scenarios. We have addressed this important problem methodically, using a novel, backplane-based approach[ICDCS'01]. With our techniques, it is now possible to use a variety of models together (e.g., combined wireless and wired networks), with disparate packages such as pdns, GlomoSim, Qualnet, GTNetS and Opnet (e.g., see [PADS'02]).
Specifically, I have worked on integrating wired/wireless network models across pdns, Qualnet and GlomoSim, resolving their semantic incompatibilities and synchronization issues. The Qualnet/Glomosim-related work was performed in collaboration with the UCLA (Rajive Bagrodia’s) group. The GlomoSim & pdns integration was in the context of a naval Future Combat Systems (FCS) scenario with SPAWAR Naval Systems Center in San Diego.
High Level Architecture
Defense Modeling & Simulation Standards, such as the HLA RTI, are sharply focused on addressing interoperability concerns. I am interested in such large, systems-of-systems approaches, especially in their systems implications such as performance and parallel scalability issues.
I was co-author of a high-performance implementation [FDK] of the DoD HLA Runtime Infrastructure Interface[DS-RT'00]. Our group’s implementation set an early standard on performance levels realizable for the HLA RTI, thereby raising the bar with respect to performance. It is one of the few “source-available” RTI implementations available, filling its place for non-commercial and academic research. Our FDK package is widely disseminated around the world. Our expertise in HLA is well-recognized, and I have been invited to present tutorials in this area to government, industry and military audiences[IITSEC'04].
We also were among the first to federate Time Warp-based simulators over an HLA-like interoperability runtime infrastructure [PADS'00].
Related Publications
- Tutorial - Parallel and Distributed Simulation and the High Level Architecture
Interservice/Industry Training, Service and Education Conference (I/ITSEC), Orlando, Florida
December 2004. [Slides]
- Experiences
Applying Parallel and Interoperable Network Simulation Techniques in On-line
Simulations of Military Networks
Kalyan Perumalla, Richard Fujimoto, Thom McLean and George Riley
ACM/IEEE/SCS Workshop on Parallel and Distributed Simulation (PADS), Washington, DC, May 2002.
[Paper][Slides]
- Distributed Network Simulations using the Dynamic Simulation Backplane
George Riley, Mostafa Ammar, Richard Fujimoto, Donghua Xu and Kalyan Perumalla
International Conference on Distributed Computing Systems, April 2001 (ICDCS), April 2001.
[Paper]
- Design of High-
performance RTI software
Richard Fujimoto, Thom McLean, Kalyan Perumalla and Ivan Tacic
Distributed Simulations and Real- time Applications (DS-RT), August 2000.
[Paper]
- An Approach to Federating
Parallel Simulators
Steve Ferenci, Kalyan Perumalla and Richard Fujimoto
ACM/IEEE/SCS Workshop on Parallel and Distributed Simulation (PADS), Bologna, Italy, May 2000.
[Paper][Slides]
- Federated Simulations Development Kit
www.cc.gatech.edu/computing/pads/fdk.html.